Interfaces with Other DisciplinesAre the least successful traders those most likely to exit the market? A survival analysis contribution to the efficient market debate
Introduction
False assumptions regarding the efficiency with which financial markets process information can lead to the misallocation of resources. A prevailing view for half a century was that prices fully discount all available information; reflected in the efficient markets hypothesis (EMH) (Fama, 1970). However, Grossman and Stiglitz (1980) challenged the existence of perfectly efficient markets, ones in which information is instantly discounted in market prices. They argued that those who buy and sell financial instruments (traders) only have an incentive to acquire costly information if profit opportunities exist. This led some to argue that the efficiency of markets may change over time (Campbell, Lo & MacKinley, 1997). More recent empirical studies confirm that the assumptions of the EMH may not hold. For example, market returns do not simply follow random walks (e.g., Doyle & Chen, 2013; Shiller, 2000) and a linear dependency between past and future prices can occur (Urquhart & Hudson, 2013). Furthermore, statistical analysis has shown that the assumptions underpinning widely used models of financial behaviour, based on the EMH, are questionable (e.g., Akerlof & Shiller, 2009; Gan, Wang Chun & Johnstone, 2017).
Psychologists and behavioural economists have also argued that the human rationality assumptions underlying the EMH do not match individual behaviour and that information overload may result in traders making poor decisions (e.g., Frino, Grant & Johnstone, 2008; Goodwin, Önkal & Thomson, 2010). This led to a shift to studying financial markets in terms of the behaviour of traders, leading to Lo (2004; 2005) arguing that markets should be viewed as a continuously evolving process; ideas encapsulated in a revised version of the EMH, the Adaptive Markets Hypothesis (AMH). Proponents of the AMH postulate that traders act in their own interests but make mistakes. As they seek to adapt, an evolutionary process based on the dynamics of competition and natural selection occurs. This results in changes in market efficiency as the populations of traders with different characteristics evolve (Urquhart & McGroarty, 2014). In particular, market forces drive those who are least likely to push prices to efficient levels under the prevailing conditions to exit the market.
The AMH suggests that “survival is ultimately the only objective that matters for all financial traders” (Lo, 2004, p25). However, most literature exploring the relationship between traders’ behaviour and survival has focused on theoretical supposition and analysis. For example, Hirshleifer (2001) postulated how judgement and decision biases incorporated in Prospect Theory (Kahneman & Tversky, 1979) may affect investors’ behaviour and asset pricing. Beyond this theoretical analysis, support for the AMH is largely based either on narrative descriptions or on market-level empirical work. Lo's (2004) work is a good example of a narrative account. He provides a description of and rationale for the AMH but offers no supporting empirical evidence. Research providing empirical support for the AMH is exclusively based on market price analysis (e.g., Ghazani & Araghi, 2014; Urquhart & McGroarty, 2014; Zhou & Lee, 2013). Despite this growing evidence in favour of the AMH, the EMH is still a widely accepted theory in the finance literature and Paul Volcker, the former Federal Reserve Chairman, claimed that "amongst the causes of the 2008–09 financial crisis was an unjustified faith in rational expectations [and] market efficiencies" (Volcker, 2011, p. 1).
This highlights the need for further research to deepen our understanding of the relative merits of AMH vs. EMH. A vital missing element is the empirical analysis of the individual trader's behaviour and how this impacts their probability of continuing to trade. This is important because a key assumption of the AMH is that market forces lead to changes in the relative proportions of ‘noise traders’, those whose decisions are based on incorrect analysis or perceptions, and traders who behave more in line with rational expectations (i.e., individuals who make decisions based on their human rationality, the information available to them, and their past experiences), referred to as ‘more informed’ traders. However, no previous studies examining the AMH (e.g., Hirshleifer, 2001; Lo, 2004; Urquhart & McGroarty, 2014) have conducted empirical analysis to explore to what extent a trader's decision to cease trading is related to the degree to which they act in accordance with rational expectations. We fill this research gap by using operational research (OR) modelling techniques, particularly survival analysis (e.g., Cox, 1972), to analyse the trades of a large data set of individuals.
The strength of OR methods for this task is that they offer the prospect of estimating mathematical models to quantify the effect of different factors on an individual's probability of ceasing to trade. This approach has been successfully employed to develop insights regarding other aspects of financial markets (e.g., Bellini & Figà-Talamanca, 2005; Doyle & Chen, 2013; Krauss, Do & Huck, 2017; Lessmann, Sung, Johnson & Ma, 2012; Moreno & Olmeda, 2007; Sermpinis, Theofilatos, Karathanasopoulos, Georgopoulos & Dunis, 2013; Tabak & Lima, 2009).
Survival analysis is the most appropriate OR method to help answer the key research question addressed here: Are noise traders (cf. more informed traders) those most likely to cease trading?’ It is widely accepted that noise traders are those most likely to display ill-disciplined, sub-optimal trading strategies and to lose money. Consequently, a methodology is employed that facilitates exploration of whether those most likely to cease trading are (i) the least profitable, and (ii) those who employ sub-optimal, ill-disciplined trading strategies.
Data related to trading in the two most popular instruments (FTSE 100 and DAX 30 futures) were provided by a large UK spread trading brokerage. These data were employed to examine the trading history of 5164 of their clients through the years 2006 to 2012. The robustness of the results was confirmed when testing the conclusions using two supplementary data sets from different spread trading brokerages based in the UK and South Africa.
To the best of our knowledge, this is the first empirical analysis of individual trader behaviour designed to test the predictions of the AMH and the first time that survival analysis has been used to study market evolution processes. Empirical evidence is presented that supports the view that changing market conditions have an impact on the mix of traders remaining in the market. In accordance with the AMH, the least profitable and the most ill-disciplined traders are found to be those more likely to cease trading than the average trader. However, the most profitable traders are also found to be more likely to cease trading than the average trader. In addition, throughout the period of the study, the ill-discipline of new generations of traders increased and during the financial crisis of 2008–09 the disposition effect of traders (the tendency to close winning positions more readily than losing positions) and the proportion of noise traders increased. We discuss how these complex new dimensions might require an adaption of the AMH.
The results are based on a specific data set drawn from one brokerage in the spread trading market and are predicated on the assumption that an individual who ceases trading with that brokerage, ceases trading completely. These facts may appear to limit the applicability of the results, but evidence is presented (see Section 3.1.2) to suggest that a spread trader who ceases trading with one broker is likely to cease trading completely. In addition, analysis of the supplementary data yielded similar results to those from the main data set, suggesting that the results have wider applicability. However, further research is needed to confirm that these results hold in other financial markets and over a longer time.
The remainder of the paper proceeds as follows. In Section 2, the literature surrounding the roles that noise traders and more informed traders may play in markets is reviewed. This literature is employed to motivate the hypotheses, which are designed to shed light on which type of trader is most likely to cease trading. In Section 3, the data set and the survival-analysis methodology employed to test the hypotheses are described. In Section 4, the results are reported and discussed. Conclusions are drawn in Section 5.
Section snippets
Hypotheses
Employing earlier theoretical and empirical research, Lo (2004, 2005) developed a modified version of the EMH (the AMH), based on the view that, like natural ecological systems, financial markets follow an evolutionary process. The AMH assumes that the behaviour of different groups of traders, distinguished by their behavioural characteristics, dictates market prices and the availability of resources (e.g., profits) and that traders who are least likely to drive prices to efficient levels
Data, variables and methodology
In this section, the nature of spread trading data is outlined, details of the specific data sets analysed are provided, and the advantages of using these data for testing the hypotheses are discussed. Second, details of the independent and control variables employed in the study are outlined. Third, the Cox Proportional Hazard Model (CPH), which was used to determine to what extent a trader's decision to cease trading is impacted by their profitability, trading ill-discipline and changing
Descriptive statistics
Descriptive statistics of the data used in the main analysis are displayed in Table 2. These relate to demographic details of the traders (e.g., age and savings) and the nature of their trading activity (e.g., staking levels, trading frequency per day, Sharpe ratios achieved). Only 25% of spread traders in the data set achieved positive average returns per trade, a similar proportion to that of profitable day traders in traditional financial markets (Barber, Lee, Liu & Odean, 2005). Whilst 61%
Conclusion
This paper tests the voracity of a key assumption of the AMH, that the dynamics of competition and natural selection will drive ill-disciplined, unprofitable traders from the market. The results indicate that such traders are not necessarily those most likely to cease trading. However, the results conform with some of the evolutionary processes that underpin the AMH, with populations of traders with different characteristics rising and falling at different times. However, the results do not
References (77)
- et al.
Good deals and benchmarks in robust portfolio selection
European Journal of Operational Research
(2016) - et al.
A survey of behavioral finance
Handbook of the Economics of Finance
(2003) - et al.
Runs tests for assessing volatility forecastability in financial time series
European Journal of Operational Research
(2005) - et al.
Exchange-rate return predictability and the adaptive markets hypothesis: Evidence from major foreign exchange rates
Journal of International Money and Finance
(2012) - et al.
Patterns in stock market movements tested as random number generators
European Journal of Operational Research
(2013) - et al.
The house money effect and local traders on the Sydney Futures Exchange
Pacific-Basin Finance Journal
(2008) - et al.
Do forecasts expressed as prediction intervals improve production-planning decisions?
European Journal of Operational Research
(2010) - et al.
On the survival of overconfident traders in a competitive securities market
Journal of Financial Markets
(2001) - et al.
Deep neural networks, gradient-boosted trees, random forests: Statistical arbitrage on the S&P 500
European Journal of Operational Research
(2017) - et al.
An application of the Cox Proportional Hazards model to bank failure
Journal of Banking and Finance
(1986)
A new methodology for generating and combining statistical forecasting models to enhance competitive event prediction
European Journal of Operational Research
Professional trader discipline and trade disposition
Journal of Financial Economics
Time is money: Costing the impact of duration misperception in market prices
European Journal of Operational Research
Information, trade and common knowledge
Journal of Economic Theory
Is the predictability of emerging and developed stock markets really exploitable?
European Journal of Operational Research
Empirical evidence on the duration of bank relationships
Journal of Financial Economics
Forecasting foreign exchange rates with adaptive neural networks using radial-basis functions and Particle Swarm Optimization
European Journal of Operational Research
Market efficiency of Brazilian exchange rate: Evidence from variance ratio statistics and technical trading rules
European Journal of Operational Research
Efficient or adaptive markets? Evidence from major stock markets using very long run historic data
International Review of Financial Analysis
Calendar effects, market conditions and the Adaptive Market Hypothesis: Evidence from long-run US data
International Review of Financial Analysis
Fundamental traders' tragedy of the commons’: Information costs and other determinants for the survival of experts and market participants in financial markets
Economic Modelling
Animal spirits: How human psychology drives the economy, and why it matters for global capitalism
Differences in risk aversion between young and older adults
Neuroscience
Boys will be boys: Gender, overconfidence, and common stock investment
Quarterly Journal of Economics
Who loses from trade?
Evidence from taiwan. working paper
Do day traders rationally learn about their ability?
SSRN Electronic Journal
The Courage of Misguided Convictions
Financial Analysts Journal
What drives the disposition effect? An analysis of a long-standing preference-based explanation
The Journal of Finance
Realization Utility
Journal of Financial Economics
Are investors really reluctant to realize their losses? Trading responses to past returns and the disposition effect
Review of Financial Studies
If you're so smart, why aren't you rich? Belief selection in complete and incomplete markets
Econometrica : journal of the Econometric Society
Market competition and selection
New palgrave dictionary of economics
White paper on spread betting
The reach of the disposition effect: Large sample evidence across investor classes
International Review of Finance
Nonlinearities in financial data
The econometrics of financial markets
Partial likelihood
Biometrika
Regression Models and Life Tables
Journal of the Royal Statistical Society, Series B (Methodological)
The Survival of market participants in financial markets
Journal of Business
Cited by (6)
A hybrid repair-replacement policy in the proportional hazards model
2022, European Journal of Operational ResearchCitation Excerpt :Examples of such systems include heave hauler trucks (Banjevic & Jardine, 2006), boring machines (Zheng, Najafi & Zhang, 2022), automotive lead-acid batteries (Man & Zhou, 2018), pumps (Zhang, Hua & Xu, 2014), and wind turbines (Zhao, Makis, Chen & Li, 2019). Cox's proportional hazards (PH) model is very popular in reliability engineering for describing the hazard function integrating the effects of both age and covariates (Blumenstock, Lessmann & Seow, 2020; Ma, Fraser-Mackenzie, Sung, Kansara & Johnson, 2021). With the development of sensor technology, condition-based maintenance (CBM) has received increasing research attention due to its efficiency and effectiveness in preventing future failures and reducing maintenance costs (Uit Het Broek, Teunter, de Jonge & Veldman, 2021; van Staden & Boute, 2021; Wang & Zhu, 2021).
Noise traders in an agent-based artificial stock market
2023, Annals of Operations ResearchResearch on Mining Balanced Competition Strategy in Financial Market Based on Computer Data Mining Method
2022, Mobile Information Systems